Software Engineer, Robot Autonomy (Actuator Control & Locomotion), Intern

The role is a Software Engineer, Robot Autonomy (Actuator Control & Locomotion) Intern at Laelaps AI. The company focuses on developing intelligent software for physical security, enabling robots and sensors to handle critical tasks that humans shouldn't. Their mission is to create a safer, more secure world through advanced robotics and AI. The main day-to-day responsibilities involve training and evaluating reinforcement learning policies for locomotion and low-level control in simulation. This includes applying techniques like domain randomization and reward design, and testing the results on physical robots. The intern will also work on making policies robust to varied terrain, challenging weather, and noisy sensor data. Additionally, the role involves exploring control approaches that can be applied across different robot embodiments and building evaluation workflows with clear metrics and repeatable experiments. To qualify for this role, candidates must be currently pursuing a PhD or have recently completed a Master's degree in Robotics, Machine Learning, Computer Science, or a closely related field. Hands-on experience training reinforcement learning policies for robot control, either in simulation or on hardware, is essential. Familiarity with robotics simulators such as Isaac Sim, MuJoCo, or Gazebo, and a solid grounding in robot dynamics and control are also required. Good coding skills in Python, with PyTorch (or JAX), and comfort using Docker and Git in workflows are must-haves. Nice-to-have skills include experience deploying learned policies on physical robots, particularly legged robots, and experience with low-level actuator, motor, or joint control. Knowledge of C++ and ROS 2, as well as publications at top robotics or ML venues, are also beneficial. The location in Zurich offers a dynamic environment with a team of PhD-level co-founders in AI, Robotics, and Physics, providing a unique opportunity for growth and learning. What the role asks for: - Currently pursuing a PhD or recently completed a Master's degree in Robotics, Machine Learning, Computer Science, or a closely related field. - Hands-on experience training reinforcement learning policies for robot control, in simulation or on hardware. - Experience with robotics simulators such as Isaac Sim, MuJoCo, or Gazebo. - Solid grounding in robot dynamics and control. - Good coding skills in Python, with PyTorch (or JAX). - Comfortable using Docker and Git in your workflows. - Experience deploying learned policies on physical robots, ideally legged (nice to have). - Experience with low-level actuator, motor, or joint control (nice to have). - Knowledge of C++ and ROS 2 (nice to have). - Publications at top robotics or ML venues (CoRL, RSS, ICRA, NeurIPS) (nice to have).